Consequences of social defeat stress for behaviour and sleep. Short-term and long-term assessments in rats

(1)

Consequences of social defeat stress for behaviour and sleep

Short-term and long-term assessments in rats

Anne Marie Kinn Rød

Dissertation for the degree philosophiae doctor (PhD) at the University of Bergen

2014

(2)

Sketch of rats in a social defeat confrontation on the front page is illustrated by Anne Marie Kinn Rød

(3)

Scientific environment

Working with the thesis, I have been employed as a PhD student at the Department of Biological and Medical Psychology, University of Bergen, Norway (2006-2011). The PhD grant was provided by the University of Bergen. I have also received a grant from the Norwegian Competence Center for Sleep Disorders, (January 2012).

The work for study I in this thesis was carried out at the Section of Physiology, Department of Biomedicine, University of Bergen, from July 2004 to August 2005, and resulted in my thesis for Master of Science. The thesis was edited to a paper which was published in 2008. The work for studies II and III was carried out at the Department of Biological and Medical Psychology, Research Group on Experimental and Clinical Stress and Sleep, University of Bergen, Norway. I was associated with The International Graduate School in Integrated Neuroscience at Department of Biological and Medical Psychology. My main supervisor was Robert Murison, and my co-supervisors were Anne Marita Milde, Janne Grønli and Håkan Sundberg, all with affiliation to the Department of Biological and Medical Psychology. Janne Grønli is also affiliated to the Norwegian Competence Center for Sleep Disorders (SOVno), Haukeland University Hospital. Anne Marita Milde is also affiliated to Resource center on Violence Traumatic stress and Suicide prevention (RVTS), Region West.

(4)

Acknowledgements

First of all I would like to express my deepest gratitude and appreciation to my main supervisor Robert Murison. Without your support, patience, shared knowledge, availability, encouragement and effort during the writing processes, I would not have been where I am today. Special thanks to my co-supervisor, Anne Marita Milde, for helping out in the lab, for helpful feedback on manuscripts, and for encouragement during rough times. My appreciation also goes to co-supervisor, Håkan Sundberg, for help with the start up of the animal lab and for guiding feedback on the thesis. And the last, but not less brilliant flower in my bouquet of supervisors, Janne Grønli, I thank for collaboration in the lab, for challenging me, for quick and to the point feedback on manuscripts, and for the opportunity to continue with sleep research. To all my supervisors, you have my deepest respect.

I wish to direct several thanks to my co-authors for their contributions. Special thanks to Jelena for collaboration in the lab (making perfect pockets), for productive discussions and for your friendship. Appreciation also goes to Finn Jellestad (luckily not the pocket man) for kind support with the startle equipment, especially the Saturday after snowfall in December 2009.

I have had great help with analyzing the corticosterone data, so thanks to Randi Espelid, Eli Nordeide and Nina Harkestad for analysing numerous blood samples. Nina, I really appreciate sharing the ‘biolab office’ with you, and I am looking forward to giving it all for ‘blood spot’. I thank Maria Maier for doing a great job on reference checking. Also, thanks to Dag Hammerborg and John Clark for saving drowning computers and for everyday computer assistance.

Credit goes to colleagues in sleep research. Especially I want to thank Chiara Maria Portas, my supervisor during my master thesis, for introducing me to the field of sleep research and together with my co-supervisor Reidun Ursin, for being a great inspiration.

A warm thank goes to my colleagues at the Department of Biological and Medical Psychology for your support, encouragement, and social ‘fredagskos’ with wine lottery. Thanks to Kenneth Hugdal for taking me under his wing and giving me

(5)

work while awaiting position announcements. Special thanks to Vivian Fosse for being the social promoter of the department, and for keeping track on all of us and all our orders. To my former colleagues Ingrid Orre and Merethe Nygård, I thank you for your friendship and for making office-sharing a pleasure. I have felt very included in the research group RECSS. Thank you for the meetings on achieved goals and shared frustrations, and thank you for the unforgettable trip to Oxford.

Many thanks go to my friends and family for their encouragements, love and support. Especially I would like to thank my parents, Marit and Morten Kinn, and my parents in law, Elsy and Terje Rød for valuable help with babysitting during the last hectic months. My dearest Erik! Thank you for being my best friend and husband, for loving me unconditionally, encouraging me, comforting me, taking care of our children and “running the household”. Without you this thesis would never have been finished. During these years as a PhD candidate, I have become what I have always wanted to be, a Mum. Siren and Kristian, I love you with all my heart!

Anne Marie Kinn Rød Bergen 21.11.2013

(6)

List of abbreviations

ACTH adrenocorticotropic hormone ANOVA analysis of variance

ASD acute stress disorder ASR acoustic startle response CMS chronic mild stress

CRH corticotropin-releasing hormone

DSM-IV Diagnostic and Statistical Manual of Mental Disorders- 4^th edition

DSM-IV-TR Diagnostic and Statistical Manual of Mental Disorders- 4^th edition-Text Revision

DSM-5 Diagnostic and Statistical Manual of Mental Disorders- 5^th edition

EEG electroencephalogram

EMG electromyogram

EOG electrooculogram EPM elevated plus maze FF fronto-frontal FP fronto-parietal

GAD generalized anxiety disorder

HPA hypothalamo-pituitary-adrenocortical

ICD-10 International statistical classification of diseases and related health problems, 10^th edition

IFS inescapable foot shock IVC individually ventilated cages LSD least significant difference MDD Major Depressive Disorder NREM non rapid eye movement

OF open field

PTSD post-traumatic stress disorder

(7)

REM rapid eye movement

sc subcutaneous

S.E.M. standard error of the mean SD social defeat

SDF social defeat fighter SDS social defeat submissive SWS slow wave sleep

SWA slow wave activity

Vmax maximum response amplitude

(8)

Abstract

Social stressors play a major role in the pathogenesis of affective disorders like anxiety and depression. These disorders are associated with altered behaviour (i.e.

locomotor activity, harm avoidance, startle response, anhedonia and sexual behaviour), sleep alterations and abnormalities in the stress response. The animal social defeat (SD) model is based on a natural conflict situation where a male intruder rat eventually subordinates itself to an unfamiliar territorial resident conspecific. The effects of defeat are studied in the intruder rat.

The main purpose was to study the face validity of the SD model for affective disorders by investigating short-term and long-term consequences of single and/or double exposure to SD on behaviour and sleep in rats. In particular, the intention was to evaluate if SD could reproduce the alterations in locomotor activity, harm avoidance, startle response, anhedonia, sexual behaviour, stress responses and sleep parallel to those observed in patients with affective disorders.

Social defeat induced low activity in the central sector of the open field (OF) (Paper I), indicating high harm avoidance which may reflect anxiety-like or depression-like behaviour. No short-term or long-term effects were seen on total locomotor activity in the OF (Papers I and II). Further, a lack of habituation to the OF across days was seen, which may reflect long-lasting heightened anxiety (Papers I and III). Overall, in the elevated plus maze (EPM) test, SD rats showed less total locomotor activity, less percentage time and less activity on the open arms, lasting up to 3 weeks after defeat (Paper II), possibly reflecting anxiety-like behaviours. High acoustic startle responses (ASR) were seen as a long-term effect of SD, probably reflecting an anxiety-like state (Paper II). A short-lasting reduced preference for sucrose was seen (Paper II), indicating an anhedonic state that may be interpreted as a transient anxiety-like symptom. Sexual behaviour was not affected (Paper I). As a group, SD rats did not show altered corticosterone responsiveness to OF exposure (Paper III).

The SD rats showed a short-term increase in duration of slow wave sleep (SWS) 2 and sleep fragmentation (Paper I). Overall, SD rats did not show long-term effects on sleep or EEG power (Paper III). The effects of SD on sleep may be

(9)

interpreted as anxiety, because they were short-lasting and the common sleep alterations seen in depression were not induced (e.g. reduced SWS2 and REM sleep alterations).

A secondary aim was to compare effects of SD to the effects of inescapable footshock (IFS) (Paper II). The two stressors induced a similar short-term effect on sucrose preference and similar long-term anxiety-like behaviours in the EPM test.

Contrary to what was expected, SD rats showed the highest ASR, while IFS rats showed the lowest total activity in the OF test. The results may reflect fundamental differences between SD and IFS.

Another secondary aim was to explore the relationship between levels of corticosterone prior to SD or IFS stressor, and the different post-stressor behaviours (Paper II). Low pre-stress corticosterone level was expected to be associated with anxiety-like behaviours following stress. Overall, such a relationship was not found.

Contrary to what was expected, the SD rats with high pre-stressor corticosterone level showed the greatest ASR, while IFS rats with low pre-stress corticosterone level did not show alterations in ASR. This further supports differences between the SD and the IFS stressor.

The final secondary aim was to investigate differences in effects of SD on behaviour and sleep in two subgroups of rats with different coping styles in the SD (Paper III). Contrary to what was expected, rats fighting back in the SD confrontation showed longer latency to leave the start box, and spent less time in the OF arena compared to those not fighting back, indicating anxiety-like behaviour. They also showed more fragmentation of sleep in SWS1 and SWS2. The results may suggest that rapid submission during SD may be more adaptive than surrender after a longer fight, given these outcome measures.

In conclusion, the studies presented in this thesis show that exposure to SD induced both short-term and long-term consequences for multiple behavioural features and at least short-term consequences for sleep. The behavioural consequences of SD are different from those of IFS.The studies generally support a high degree of face value for the SD model as a model for affective disorders, more relevant to anxiety than to depression.

(10)

List of publications

This thesis is based on the following papers.

Paper I. Kinn AM, Grønli J, Fiske E, Kuipers S, Ursin R, Murison R, Portas CM.

2008. A double exposure to social defeat induces sub-chronic effects on sleep and open field behaviour in rats. Physiology & Behavior, 95(4), 553- 561.

Paper II. Kinn Rød AM, Milde AM, Grønli J, Jellestad FK, Sundberg H, Murison R.

2012. Long-term effects of footshock and social defeat on anxiety-like behaviours in rats: Relationships to pre-stressor plasma corticosterone concentration. Stress 15(6), 658-670.

Paper III. Kinn Rød AM, Murison R, Mrdalj J, Milde AM, Jellestad FK, Øvernes LA, Grønli J. Effects of social defeat on sleep and behaviour: importance of the confrontational behaviour. Physiology & Behavior

(resubmitted manuscript, PHB-D-13-00281R1)

(11)

Chapter 1 Introduction

1.1 General introduction

Social stressors are one of the main sources of stress in human life, especially for those low in the social hierarchy (Wood AM et al. 2012), and play a major role in the pathogenesis of affective disorders i.e. anxiety and depression (Taylor et al.

2011). Social stress can occur throughout the lifespan, from childhood neglect, abuse, and school bullying to work harassment in adulthood (Bjorkqvist 2001; Heim and Nemeroff 2001), or may be associated with traumatic events like violence and assault (Krug et al. 2002).

The interest for animal models of social stress has recently increased, especially the social defeat (SD) model, possibly due to the recognition that social stress is highly associated with pathology and the acknowledgement that natural stress models have important translational value (Chaouloff 2013). The SD model is based on the natural conflict occurring when a male intruder rat (or mouse) eventually subordinates itself to an unfamiliar territorial resident conspecific. The effect of defeat is studied in the intruder rat, and may induce short-lasting and long- lasting alterations in behaviour and physiology.

1.2 The concept of stress

Living organisms have a complex set of mechanisms to maintain constancy of their internal environment and to preserve life. Our understanding of these

(16)

mechanisms started with Claude Bernard’s concept of an internal environment, milieu interne (Bernard 1885). This was further elaborated by Walter Cannon (Cannon 1932) who coined the term homeostasis, meaning steady state. Homeostasis is the maintenance of a relatively constant internal environment by an array of mechanisms in the body. Cannon recognized that emotional as well as physiological disturbances activate a sympathoadrenomedullary response, the ‘fight or flight’ response, preparing the body for action. The concept of stress was first used in the biomedical literature by Hans Selye in the 1930s. Selye outlined the ‘general adaptation syndrome’, the consistent sequence of three stages of physical responses triggered by a stressor (Selye 1936).

Stress has been defined as a state in which homeostasis is threatened or perceived to be threatened (Chrousos and Gold 1992). Stressors are the physical or psychological stimuli which threaten homeostasis. Later, the term allostasis has been used in the literature, defined as the processes actively maintaining homeostasis (McEwen 2000, 2010). The hypothalamo-pituitary-adrenocortical (HPA)-axis and the sympathoadrenomedullary system are such adaptive processes and promote adaptation and coping. When the body is forced to adapt to adverse situations, the cost to the body is allostatic load (McEwen 2000, 2010). Frequent stress, failure to habituate to repeated challenges, inability to shut off responses and inadequate responses are conditions which may lead to allostatic overload, which consequently may cause pathology.

The definition of stress has been debated because nearly all activities of an organism directly or indirectly threaten homeostasis, and the stress response is also activated during rewarding behaviour like sexual behaviour and winning a social interaction (Buwalda et al. 2012). In a recent review of the stress concept, Koolhaas and colleagues (2011) emphasized that:

The use of the terms ‘stress’ and ‘stressor’ should be restricted to conditions and stimuli where predictability and

controllability are at stake; unpredictability being characterized by the absence of an anticipatory response and

(17)

loss of control being reflected by a delayed recovery of the response and the presence of a typical neuroendocrine

profile. (p. 1292).

This definition of stress and stressors exclude the short lasting adaptive activation of the stress response (Koolhaas et al. 2011).

Stress is thought to be maladaptive and potentially pathogenic if the response is sustained and not adequately terminated (Ursin H and Eriksen 2010). It has been emphasized that stress should be considered as a process that includes the stimulus, the perceptual processing of the input and the behavioural and physiological output (response) (Levine 2005). This approach forms the basis of the cognitive activation theory of stress (CATS: Ursin H and Eriksen 2010).

Stress is hypothesized to induce a cascade of behavioural and neurobiological processes, with possible different time-courses for each process (Koolhaas et al.

1997b). Some of these processes return to baseline after a few hours, others take days or weeks, and possibly some of the processes are indefinitely changed, never to return to baseline (Koolhaas et al. 1997b). The different temporal dynamics of the various stress parameters imply that the physiological and behavioural state of the individual at one time-point after stress is different from its state at a later time-point.

Additionally, the vulnerability to subsequent stressors may vary with the state of the individual at different points in time. Thus, the symptomatology will be different depending on the time of measurement after stress and on subsequent stressors.

1.3 The stress response system

The main components of the stress response system are the central nervous system, the peripheral nervous system consisting of the autonomic and the somatic nervous system, and the HPA-axis (for reviews see e.g. Chrousos 2009; Chrousos and Gold 1992; Vermetten and Bremner 2002).

When a threat or aversive event is registered by the central nervous system, either as an environmental stimulus or as a memory of the previous aversive experience, the immediate response involves increased sympathetic activation via the

(18)

autonomic nervous system (within seconds, Eriksen et al. 1999), which again leads to responses that are essential to prepare the body for fight or flight. Peripheral sympathetic activation increases heart rate, blood pressure, respiration and metabolism. Noradrenergic activation throughout the brain leads to enhanced arousal, vigilance, focused attention and increased activity of the HPA-axis. Additionally, activation of the amygdala by noradrenaline is important for memory-retrieval and emotional analysis of the stressor (McGaugh 2000). If the stressor is of a threatening nature, the amygdala activates the stress system (LeDoux 1994). Also, sympathetic activation of the adrenal medulla leads to release of adrenaline and noradrenaline, which have partly the same effects in the body as sympathetic activation. The parasympathetic arm of the autonomic nervous system counteracts the sympathetic arm to prevent an exaggerated response. The two arms of the autonomic nervous system are at all times activated, but the balance of activation leads to specific responses, like fight/flight, sleep/wakefulness, digestion, reproduction, and so forth (Chrousos and Gold 1992; McCarley 2004; Van Reeth et al. 2000; Vermetten and Bremner 2002).

The other and slower part of the stress response system is the HPA-axis.

Neuroendocrine cells in the hypothalamus release corticotropin-releasing hormone (CRH) into the blood vessels surrounding the pituitary stalk, leading to the release of adrenocorticotropic hormone (ACTH) from the anterior lobe of the pituitary into the blood (within seconds, Eriksen et al. 1999). Cells in the adrenal cortex are stimulated by ACTH to release glucocorticoids into the bloodstream (within minutes, Eriksen et al. 1999). The main glucocorticoid is cortisol in humans and corticosterone in rats.

Glucocorticoids have a negative feedback effect on the HPA-axis by down-regulating the release of CRH and ACTH, both directly and indirectly via the hippocampus. In addition to the increased release of glucocorticoids as a response to stress, glucocorticoids (like essentially all hormones) are released in a circadian pattern in both rats (Allen-Rowlands et al. 1980) and humans (Weitzman et al. 1971).

The complexity of the glucocorticoids is illustrated by its effect on behaviour, arousal, sleep, brain development and function, and bodily functions like the immune

(19)

system, endocrine systems and energy mobilisation, (Chrousos and Gold 1992;

Lupien et al. 2009; Van Reeth et al. 2000).

1.4 The basics of sleep and sleep regulation

Exposure to stress may lead to altered behaviour and sleep. This is seen in human affective disorders that are associated with stress, and will be introduced below (1.5 Affective disorders associated with stress). Because the consequences on sleep are an important part of the thesis, the basics of sleep and sleep regulation will be described in some detail in the following.

Normal sleep in humans and animals comprises a complex combination of physiological and behavioural processes. Sleep may be defined as a reversible behavioural state of partial perceptual disengagement from and unresponsiveness to the environment (Carskadon and Dement 2011). However, the sleeping brain can be easily woken when given a sufficient level of stimulation, and to some degree it distinguishes important information from unimportant information, a crucial feature which makes it possible to wake up when danger is present (Portas et al. 2000).

Sleep and wakefulness in both humans and animals are objectively measured by polysomnography, a set of electrophysiological parameters where the core measures are the electroencephalogram (EEG – brain activity), electromyogram (EMG – skeletal muscle activity) and electrooculogram (EOG – eye movements, not often used in rats). On the basis of these parameters, the different sleep stages can be defined and quantified by a set of scoring criteria in humans (Rechtschaffen and Kales 1968) and rats (Neckelmann and Ursin 1993; Ursin R and Larsen 1983).

Normal sleep is divided into two main phases in all mammals: non rapid eye movement (NREM) sleep and rapid eye movement (REM) sleep. Through the inactive phase, sleep progresses in cyclic patterns. One of these is the alternation between NREM and REM sleep, which has 4-5 cycles throughout the night in humans. Sleep starts in NREM and progresses through the sleep stages of NREM before entering the first REM sleep episode after about 90 minutes in humans (Carskadon and Dement 2011), and 12 minutes in rats (McCarley 2007).

Characteristic for the NREM sleep in both humans and rodents is reduced EMG

(20)

activity and the presence of sleep spindles and high-voltage slow waves (delta waves) in the EEG. In humans, NREM is classically subdivided into stages 1-4 (Rechtschaffen and Kales 1968). Stages 3 and 4 contain the highest intensity of slow waves (high total power, amplitude, and incidence of slow waves), and is collectively named deep slow wave sleep (SWS). In rats, NREM sleep is subdivided into SWS1 and SWS2 (Ursin R and Larsen 1983), where SWS2 contains the highest amount of slow waves, and is comparable to deep SWS in humans. REM sleep is characterized by the occurrence of rapid eye movements shown in the EOG, desynchronized EEG similar to wakefulness, and very low EMG activity (atonia) with occasional muscle twitches. In rats, low EMG activity and the presence of theta activity obtained by intracranial EEG gives adequate characteristics for the REM sleep. Wakefulness is characterized by desynchronized EEG and high EMG activity. Visual scoring of the sleep recording provides detailed information of sleep pattern, for example sleep latency, total sleep time, time spent in each sleep stage, number of stage shifts (fragmentation) and number of awakenings. These parameters represent a good measure of sleep quality and quantity.

Information of EEG power during sleep may be obtained from power spectral analyses (e.g. fast Fourier transform), which describe power or energy distribution in each EEG frequency band. Power values reflect both the incidence and amplitude of waves. Slow wave activity (SWA) is the EEG power in the low-frequency/delta band (e.g. 0.5-4.5 Hz, a definition which may vary between studies in both humans and rats). Compared with wakefulness, the overall power of the EEG increases during sleep, reflecting greater synchrony of CNS activity, and is greatest during deep SWS (Greene and Frank 2010).

The timing and quality of sleep is regulated by the interaction between sleep need (homeostatic factor), circadian factors and behaviour (Ursin R 2008). The homeostatic factor accumulates during time spent awake (Borbely 1982), and is reflected in the amount and intensity of the deep SWS (Achermann and Borbely 2003). The circadian factor (mediated by the suprachiasmatic nuclei) promotes sleep during certain periods of the day, and determines to a large degree the timing and duration of the sleep period (Dijk and von Schantz 2005). In addition to the

(21)

homeostatic factor and the circadian factor, deactivation by behaviour is necessary.

Lying down in a safe environment facilitates muscle relaxation and reduced activation of the brain. Exposure to a stressful stimuli leads to activation of the stress system and stimulates arousal, which again supresses and alters sleep (Chrousos 2009; Van Reeth et al. 2000). Stress is thus a state where sleep homeostasis and sleep behaviour are threatened, and may induce changes in sleep regulation.

1.5 Affective disorders associated with stress

Although stress is naturally occurring and induces an adaptive response, it is also believed to play a major role in the pathogenesis of affective disorders. While most, if not all, people experience severe stress in the course of their lives, only a minority will develop a disorder (Kessler et al. 1995), apparently reflecting an abnormal response to stress rather than the norm. The reasons why some develop disorders, while others do not, may be explained by the diathesis stress hypothesis, which proposes that the interaction of diathesis (predisposition or vulnerability) to affective disorders and the experience of stressful events may result in psychopathology. The predisposition or vulnerability can involve e.g. a particular genetic makeup, physiology or personality, or a combination of these. In addition to these, the characteristics of the stressor or trauma play a role in development of a disorder or resistance/resilience to it (Monroe and Simons 1991). Anxiety and depression are complex affective disorders associated with stress.

Criteria for the classification of affective disorders and other mental disorders have been developed to provide guidance to clinicians and researchers, for example The Diagnostic and Statistical Manual of Mental Disorders 4^th edition (DSM-IV, American Psychiatric Association 1994), 4^th edition-Text Revision (DSM-IV-TR, APA 2000), 5^th edition (DSM-5, APA 2013), and the International statistical classification of diseases and related health problems (ICD-10, World Health Organization 2008).

1.5.1 Anxiety disorders

According to DSM-IV-TR (APA 2000) anxiety disorders are classified as phobias, panic disorder, obsessive-compulsive-disorder (OCD), generalized anxiety

(22)

disorder (GAD), acute stress disorder (ASD) and post-traumatic stress disorder (PTSD). Common for these is autonomic activation which induces symptoms like increased heart rate, high blood pressure, respiratory changes, altered metabolism, enhanced arousal, vigilance and focused attention. In the USA, the prevalence for anxiety disorders is reported to be for lifetime 28.8% and for 12-months 18.1%

(Kessler et al. 2005a; Kessler et al. 2005b). Animal models of anxiety normally aim to mimic symptoms of GAD, ASD and PTSD.

Fear versus anxiety

Fear and anxiety are emotional states that mediate survival responses to threats (Porges 1995). In neuroscience, fear is commonly defined as an aversive reaction elicited by the perception of a specific threat stimulus, whether conditioned or unconditioned. Anxiety, in contrast, is commonly defined as prolonged hyper- vigilance in anticipation of, or response to, a diffuse or imagined threat where danger is not clearly imminent or not present (Sylvers et al. 2011). The states are similar as autonomic arousal occurs in both, but the relationship to HPA-activation is less clear for anxiety than it is for depression. There are also other important distinctions between fear and anxiety. The fear response dissipates quickly, whereas anxiety promotes a sustained response. The two responses are mediated by different brain regions: the central amygdala is the primary brain structure in fear, whereas the bed nucleus of stria terminalis is the primary brain structure in anxiety (Sylvers et al.

2011). Anxiety disorders as used in the diagnostic classification may thus be seen to include disorders of fear (e.g. phobias, social anxiety) and of anxiety (GAD, ASR, PTSD).

1.5.2 Depressive disorders

Mood disorders include conditions such as major depressive disorder (MDD), bipolar disorder, and dysthymic disorder (chronic mild depression) (DSM-IV-TR, APA 2000). Common for the mood disorders are disturbed mood as the predominant feature, and symptoms of altered behaviour and sleep. The prevalence of mood disorders is reported to be for lifetime 20.8% and for 12-months 9.5% (Kessler et al.

2005a; Kessler et al. 2005b). The most debilitating form of depression is MDD, and it

(23)

is the symptoms of MDD that most animal models of depression normally aim to mimic.

1.5.3 Altered behaviour associated with affective disorders

Corticotropin-releasing hormone has been hypothesized to have direct behavioural effects in the brain that lead to increased arousal, alertness, attention, and readiness (Vermetten and Bremner 2002). Abnormal levels of arousal are seen in both anxiety and depression. Depressed patients may show either hypo- or hyperarousal in the central nervous system (Nofzinger et al. 2000), while anxiety patients show hyperarousal (DSM-IV-TR, APA 2000). The clinical links to abnormal arousal include hypersomnia or insomnia (see 1.5.4 Sleep alterations associated with affective disorders below), increased or decreased psychomotor activity and increased startle response (DSM-IV-TR, APA 2000; Nofzinger et al. 2000).

Increased startle response is one of the diagnostic symptoms linked to increased arousal in ASD and PTSD (DSM-IV-TR, APA 2000). In GAD patients, negative emotionality is associated with an excessive startle response (Ray WJ et al.

2009). In PTSD patients, a lack of habituation of the startle response magnitude and skin conductance has been reported (Jovanovic et al. 2009; Metzger et al. 1999).

The temperament and character inventory (TCI) (Cloninger et al. 1993) and the earlier tridimensional personality questionnaire (Cloninger 1987) are questionnaires developed to evaluate the psychological and biological tendencies of human behaviour, and are widely used in psychiatry and psychology. Harm avoidance is one of the dimensions of temperament, an inherited personal trait which is stable over time. When exposed to potential threat, people with high scores on harm avoidance show caution and careful planning. Harm avoidance has been shown to positively correlate with symptoms of both anxiety and depression (Jylha and Isometsa 2006), and to correlate negatively with resilience (Kim et al. 2013). Patients with unipolar depression and PTSD show high harm avoidance (Jakšić et al. 2012;

Young et al. 1995).

One of the core symptoms of MDD is anhedonia, defined as inability to find pleasure in things usually found enjoyable (DSM-IV-TR, APA 2000), e.g.

recreational activities, eating, social interaction and sexual activities. Anxiety and

(24)

other neuropsychiatric disorders have also been associated with anhedonia (Der- Avakian and Markou 2012; Grillo 2012). Decreased sexual desire and increased sexual dysfunction are common symptoms with depression and anxiety disorders (Kendurkar and Kaur 2008; Kotler et al. 2000; Laurent and Simons 2009; Michael and O'Keane 2000), but may also be a result of the treatment (Fossey and Hamner 1994; Williams and Reynolds 2006).

1.5.4 Sleep alterations associated with affective disorders

Sleep alterations are among the symptoms for both anxiety and depression (DSM-IV-TR, APA 2000). In patients with GAD, ASD and PTSD, these subjectively reported sleep alterations are difficulty with initiation and maintenance of sleep (insomnia), which may be linked to increased arousal. Another common symptom in PTSD is the occurrence of distressing dreams about the triggering traumatic event.

For MDD the reported symptoms are insomnia, hypersomnia (daytime sleepiness) or decreased need for sleep. In addition to these subjective alterations, objective alterations have been reported in several affective disorders.

There are few studies on objective sleep alterations in GAD patients (Monti and Monti 2000). Studies have reported increased sleep onset latency, increased wake time after initial sleep onset, lower sleep efficiency, and reduced total sleep time relative to controls. Findings of abnormalities in the amount and timing of REM sleep and the amount of SWS are inconsistent in GAD patients (Monti and Monti 2000;

Papadimitriou and Linkowski 2005).

Objective findings on sleep disturbances in PTSD are also inconsistent, even the subjective reports of trouble initiating and maintaining sleep are inconsistently found in objective assessment of sleep. Reports of frequent nightmares in PTSD (which most typically arise during REM sleep) have focused interest on REM sleep alterations. Abnormalities in the timing or amount of REM sleep in PTSD have not been consistently found. However, increased REM density (frequency of rapid eye movements) and REM sleep fragmentation (arousals and stage shifts) have been reported (Kobayashi et al. 2007; Ramsawh et al. 2011).

Regarding objective sleep alterations, MDD is the most studied affective disorder. These include increased sleep onset latency, reduced total sleep time, lower

(25)

sleep efficiency, reduced REM sleep latency, increased amount of REM sleep, increased REM density, reduced amount of sleep stage 3 and 4 (deep SWS) and increased sleep fragmentation (Peterson and Benca 2011).

Changes in EEG power during sleep have also been reported in both depression and anxiety disorders, including reduced power in the low-frequency delta band (0.2-4 Hz) and increased high frequency power (>20 Hz) (Armitage 2007;

Borbely et al. 1984; Tekell et al. 2005; Woodward et al. 2000).

1.5.5 Abnormalities in the stress response associated with affective disorders

Depression in humans is often associated with higher than normal basal levels of glucocorticoids, hypercortisolism (de Kloet et al. 2005; de Villiers et al. 1987).

One possible mechanism underlying hypercortisolism is a reduced inhibition of the HPA-axis by the hippocampus. Decreased numbers of mineralocorticoid receptors and glucocorticoid receptors in the hippocampus, as seen for instance in adolescent and adult rats exposed to prenatal stress, weaken the inhibition of the stress response, resulting in increased basal and/or stress-induced glucocorticoid secretion (Lupien et al. 2009).

Post-traumatic stress disorder (PTSD) is often associated with low basal levels of glucocorticoids, hypocortisolism (Heim et al. 2000; Mason et al. 1986). However, hypocortisolism in PTSD is not a consistent finding (Eckart et al. 2009; Inslicht et al.

2006), although a meta-analysis did yield some evidence for hypocortisolism in a subgroup of people who seem to be at the greatest risk of developing PTSD (Meewisse et al. 2007). One proposed mechanism for the development of hypocortisolism in PTSD is increased CRH release, leading to a lower ACTH response to CRH, and resultant low levels of peripheral cortisol (Vermetten and Bremner 2002). The hypocortisolism in PTSD may be a pre-traumatic risk or vulnerability factor that is induced by genetic predisposition and/or early exposure to stress rather than a consequence of trauma (Pitman 1997), as suggested by the diathesis stress hypothesis.

(26)

1.5.6 Comorbidity

Among patients who meet criteria for major depression, 51% are also suffering from an anxiety disorder (Kessler et al. 1996). As described there is an overlap of diagnostic criteria and symptoms of anxiety and depression. Consequently, it is difficult to separate the disorders in humans, and thus also in animal models. There are a few possible explanations for this frequent comorbidity. One is that there may be a common underlying genetic and/or environmental factor predisposing to both conditions, which may manifest itself as anxiety or depression or both at different times in life. For instance, compared to individuals with two copies of the long allele, individuals with one or two copies of the short allele of the serotonin transporter promoter polymorphism exhibit more depressive symptoms, diagnosable depression and suicidality after exposure to stressful life events (Caspi et al. 2003), and have a higher risk of developing PTSD after adult traumatic events and childhood adversity (Xie P et al. 2009). Another explanation is that anxiety disorders may cause or contribute to the development of depression. This explanation is supported by reports that the age of onset for anxiety disorders is lower than for depression (Schatzberg et al. 1998). Additionally, life time MDD has been shown to be secondary to other mental disorders, whereas anxiety is the most common pre-existing disorder (Kessler et al. 1996).

1.6 Animal models of affective disorders

Today there is an increasing focus on affective disorders, how they impair the quality of life for the patients and how they impact on societal economies. Laboratory animal models can contribute to the understanding of the triggering environmental factors as well as the mechanisms, neurobiology and genetics behind the disorders.

These models and knowledge from them can be used to improve prediction and treatment of the disorders.

Animal models are based on the foundation that all vertebrate animals, especially mammals, have through evolution developed substantial commonalities of structure and function, from gross anatomy to organ systems and the most elemental processes between and within cells. These commonalities imply that the brain and its

(27)

regulation of behaviour in any mammal will probably have substantial generalizability to all mammals, including humans (Overmier and Carroll 2001).

Criteria have been established for the evaluation of animal models of affective disorders. Widely quoted are the criteria for models of depression developed by McKinney and Bunney (1969), which are also utilized in models of other affective disorders. They proposed that the minimum requirements for an animal model of an affective disorder are: 1) It is ‘reasonably analogous’ to the human disorder in its manifestations or symptomatology; 2) there is a behavioural change that can be monitored objectively; 3) the behavioural changes observed should be reversed by the same treatment modalities that are effective in humans; 4) it should be reproducible between investigators.

McKinney and Bunney’s criteria have later been further elaborated to account for etiological, face, predictive and construct validity (Henn and Vollmayr 2005;

Willner 1984). An animal model should have similar causative conditions to the human disorder, etiological validity; similar manifestations and symptom profiles to the disorder state, face validity; similar treatment responses to that seen in the human disorder, predictive validity; and similar underlying neurochemical processes responsible for the symptoms observed in the human disorder, construct validity.

Diagnostic criteria for affective disorders like PTSD and MDD include symptom persistence over time, such as several weeks (DSM-IV-TR, APA 2000).

Furthermore, symptoms may take time to manifest themselves following a precipitating event, e.g. delayed onset of PTSD. Thus it has been proposed that animal models for these disorders should show long-lasting changes (Stam 2007;

Yehuda and Antelman 1993).

Severe stress is a common risk factor for affective disorders like PTSD and depression (Neria and Bromet 2000). Several animal models of affective disorders are therefore based on various forms of stress (etiological validity) with the aim to induce alterations of behaviour and physiology analogous to the human disorder (face validity). Some animal models are proposed to be specific for anxiety or depression.

Others may model both anxiety and depression, which is not surprising as many symptoms of anxiety and depression overlap. Several animal models based on

(28)

environmental stress have been developed (see 1.6.2 Stress exposure as animal models of affective disorders below), and for each model a variety of protocols are used. One must have this in mind when studies are compared and lines are drawn to the human condition.

Patients with a given affective disorder have a set of behavioural and physiological symptoms characteristic for the disorder. They are diagnosed from these symptoms in addition to subjective verbally expressed symptoms. An animal model of the affective disorder aims to reproduce parts or all of the objective behavioural and physiological changes present in patients with the disorder, but cannot access subjective symptoms.

1.6.1 How to measure the face validity of animal models?

Several methods may be used to evaluate behavioural and physiological effects of animal models associated with anxiety and depression in humans, i.e. the face validity.

Over the years, a large number of tests of animal behaviour have been developed and validated as tests for anxiety-like or depression-like behaviour (Lister 1990; Overstreet 2012; Ramos and Mormède 1998). A test is said to be valid as a test for anxiety-like or depression-like behaviour if the effect is reduced by anxiolytic or antidepressant drugs, respectively. In the following there will be given an introduction to the behavioural tests used in this project.

To study the effect of sleep in rodents, measures of brain and muscle activity (EEG and EMG) are used. Sleep registration in rodents will be introduced after the introduction of behavioural tests.

Open field (OF)

The OF test consists of the measurement of behaviours elicited by placing the animal in a novel open space where escape is prevented by surrounding walls.

Several variations of the apparatus and the protocol have emerged. The two main paradigms of the test are the forced exploration OF test, where the animal is placed directly in the arena, and the free exploration OF test, sometimes named the OF

(29)

emergence test, where the animal is allowed to explore freely from a start box or home cage.

A novel arena of any sort is likely to evoke complex, competing behavioural tendencies reflecting anxiety and fear (harm avoidance) on the one hand and exploration and curiosity (novelty seeking) on the other (Ray J and Hansen 2004).

The original view was that a novel, potentially dangerous environment initiates a stress response leading to low locomotor activity and high defecation rate (Archer 1973; Denenberg 1969), an indication of increased sympathetic activity (Sapolsky 1998). Another measure used is activity or time spent in the central area. In a novel OF, the rats tend to move mostly in the peripheral area, where they can touch the walls, thereby avoiding the open, more aversive and potentially dangerous central area. An additional parameter used in the free exploration OF is latency to leave the start box/home cage, where long latency may reflect high anxiety.

Overall, the first exposure to the OF is more anxiety-provoking because of the novelty of the arena. Over repeated exposures, the field loses its novelty and habituation normally occurs, indicated by e.g. increased locomotor activity and decreased defecation (Archer 1973; Denenberg 1969). A lack of habituation in the OF may indicate a sustained state of anxiety.

Open field behaviour seen in rodents after exposure to stress parallels in many ways behaviours seen in humans with anxiety and/or depression. Avoiding the central area or high latency to leave the start box/home cage may be linked to the concept of harm avoidance (Jylha and Isometsa 2006; Ray J and Hansen 2004; Vermetten and Bremner 2002). Reduced total locomotor activity can however be interpreted as freezing behaviour reflecting fear/anxiety, or as psychomotor retardation reflecting a symptom in human depression. Increased locomotor activity may reflect the depression-like symptom of psychomotor agitation. Thus the OF test may test both anxiety-like and depression-like behaviour.

Stress-induced alterations in behaviour in the OF are reduced by some but not all anxiety reducing drugs (Prut and Belzung 2003). Additionally, behavioural alterations in the OF test following a stressor have been shown to be reversed by antidepressant drugs and by sleep deprivation (Katz et al. 1981; Meerlo et al. 1996a),

(30)

which has an acute antidepressant effect in humans (Wu and Bunney 1990). These effects on OF behaviour also indicate that the OF test may test both anxiety-like and depression-like behaviour.

Elevated plus maze (EPM)

The basis for the development of the EPM test was that rats display higher avoidance and lower exploratory behaviour in open elevated alleys compared to closed alleys (Montgomery 1955). The apparatus described by Pellow and colleagues (1985) consisted of four elevated arms, arranged in a plus shaped cross with two open and two enclosed opposing arms, connected by a central platform giving free access to all four arms. The rat is placed on the central platform and is allowed to explore the maze for a fixed amount of time. The most used parameters are related to entry, activity and duration on open arms. As in the OF test, repeated exposure to the EPM arena is likely to induce habituation, and a lack of habituation may indicate a sustained state of anxiety. Compared to the OF, the EPM apparatus is more standardized.

The EPM test has been validated by Pellow et al (1985) who found that rats consistently avoided the open arms and preferred the closed arms. Open arm approach was increased by anxiolytics, decreased by anxiogenic substances, and was unaffected by antidepressants. Thus, avoidance and lower exploration of open arms in the EPM is taken as an indicator of anxiety. As for central activity in the OF test, decreased exploration of the open arms may be interpreted as harm avoidance seen in human anxiety (Jylha and Isometsa 2006; Ray J and Hansen 2004). Taken together, the EPM test is viewed as a test of anxiety-like behaviour.

Acoustic startle response (ASR)

Startle is regarded as a preparatory reflexive behaviour. The ASR is enhanced in threatening situations or following an aversive event. In response to a loud noise, both animals and humans show a startle reflex by blinking the eyes and contracting skeletal muscles. In the ASR test the animal is put in a pressure sensitive tube inside a sound attenuated chamber. A series of intensive, sudden acoustic stimuli is presented to the animal and the magnitude of the muscular outcome of the ASR is measured by changes in pressure to the floor in the tube.

(31)

The startle response has a short latency (e.g. 8 milliseconds measured by EMG in the hind leg) and is thought to be mediated by a relatively simple neural pathway.

In rats, the most accepted primary acoustic startle reflex pathway involves three central synapses: a) auditory nerve fibres to cochlear root neurons, b) cochlear root neuron axons to cells in the nucleus reticularis pontis caudalis, and c) nucleus reticularis pontis caudalis axons to motor neurons in the facial motor nucleus (pinna reflex) or in the spinal cord (whole body startle) (Davis 2006; Koch 1999).

In humans, the startle response has greater magnitude during negative affective states (Lang et al. 1998), and exaggerated startle response is among the DSM-IV-TR (APA 2000) diagnostic criteria for PTSD and ASD. Animal studies have shown sensitization of the ASR after footshock (enhancement of the response) (Davis 1989;

Milde et al. 2003).

The sucrose preference test

When given a choice, rats as well as humans normally prefer to drink sweetened liquids. Katz (1982) reported that sucrose and saccharine consumption were reduced by chronic severe stress, indicating anhedonia, a depression-like symptom in rats. To test this stress-induced anhedonia in rats there are now several versions of this test. Animals are normally adapted to the sweet solution before the test. In one version of the test, rats are given a bottle of sweet solution for a short period (e.g. one hour). In the sucrose preference test rats are given a free choice between water and sucrose over several hours (usually 24 hours). Both sucrose and saccharin, a non-caloric artificial sweetener, have been used to test anhedonia. The use of sucrose in the test for anhedonia is a debated topic because the underlying motivation for sucrose intake may be both for caloric intake and for hedonic reasons, mediated by separate circuits in the brain (Bear et al. 2001). However, sucrose and saccharine preference tests have to a large extent been used interchangeably.

Reduced intake of a sweet solution is proposed to reflect decreased motivation and anhedonia, which is a core symptoms of depression (DSM-IV-TR, APA 2000), and is also associated with anxiety (Der-Avakian and Markou 2012; Grillo 2012).

The sucrose preference test is normally viewed as a test of depression-like behaviour, however the test cannot be ruled out as a test of anxiety-like behaviour.

(32)

Sexual behaviour

Stress leads to a decrease in reproductive hormones and altered neurotransmission underlying reproductive behaviour. Stress may also lead to anhedonia and decreased motivation which may reduce sexual behaviour. In the sexual behaviour test, a male rat is introduced to a female rat in oestrus for a pre- defined amount of time, and sexual behaviour is scored in the gender of interest.

Prolonged latency and decreased number of mounts, intromission, and ejaculation are considered signs of sexual dysfunction and decreased sexual motivation in male rodents (Argiolas et al. 1988; Hawley et al. 2011).

Humans diagnosed with anxiety and/or depressive disorders may report a reduction in libido and are at greater risk of experiencing physiological impairment in sexual functioning (e.g. ejaculatory and erectile dysfunction) (Kendurkar and Kaur 2008; Kotler et al. 2000; Laurent and Simons 2009; Michael and O'Keane 2000).

Reduced motivation for sexual activities is, like reduced sucrose preference, a sign of anhedonia (Gorwood 2008). Changes in sexual behaviour in rats may thus reflect both anxiety-like and depression-like behaviour and physiology.

Sleep registration in rodents

The objective sleep alterations seen in anxiety and depression may also be seen in rats after exposure to stress. Electrodes for EEG recording are surgically implanted on the skull of the rat, and EMG electrodes are implanted in the neck muscle. Data recordings are made by connecting the rat to a freely moving cable and the recording equipment or by wireless transmission of the data from a surgically implanted device (see 2.14.2 Sleep recording procedures). Sleep and wakefulness is scored manually according to a set of scoring criteria (e.g. Neckelmann and Ursin 1993; Ursin R and Larsen 1983), by automatic scoring algorithms, or by a combination of the manual and automatic method.

Some of the objective sleep alterations seen in patients with anxiety and depression have been shown in animals exposed to early life stress, CMS and learned helplessness (Adrien et al. 1991; Dugovic et al. 1999; Grønli et al. 2004; Mrdalj et al.

2013).

(33)

1.6.2 Stress exposure as animal models of affective disorders

Widely used stressors in animal models of affective disorders are uncontrollable and unpredictable electrical footshocks.

The learned helplessness model was originally introduced by Overmier and Seligman (1967), and was based on the observation that following a high number of repeated inescapable shocks (64 shocks in the original paper), animals (originally dogs) will not try to escape from a situation even if it is possible. Animals which had previously learned to escape shock did not develop learned helplessness when exposed to inescapable footshocks followed by escapable shock (Seligman and Maier 1967). This is thought to parallel the attitude in depressed humans, that behaviour does not influence what happens next (Miller and Seligman 1975).

The learned helplessness model is known to produce depression-like symptoms in rodents e.g. agitated motor behaviour, REM sleep alterations, reduced body weight, diminished sexual behaviour, reduced intake of sweet solution and elevated corticosterone and CRH levels (Nestler et al. 2002; Vollmayr and Henn 2003). Antidepressant drug treatment, electroconvulsive shocks and cognitive training reverse learned helplessness and the depression-like symptoms, while anxiolytics drugs do not (Nestler et al. 2002; Seligman and Maier 1967; Sherman et al. 1982). Thus, the learned helplessness model is primarily known as an animal model of depression. The model is also suggested as an animal model of PTSD (Foa et al. 1992; Krystal et al. 1989; LoLordo and Overmier 2011), although this is questioned (Yehuda and Antelman 1993).

Another model using footshock is the brief inescapable footshock (IFS) model.

In this model animals are exposed to a relatively short-lasting session with a low number of IFSs. The model was first described by Levine and colleagues (1973).

Murison and Overmier (1998) showed that there was a qualitative difference between 10 shocks and 100 shocks delivered to rats. Rats exposed to 10 shocks showed anxiety-like behaviour (immobility) in the sudden silence test, while 100 shocks had no effect. Overall, the model is known to produce anxiety-like symptoms in rats, as shocked rats show lower activity in an OF (Van Dijken et al. 1992c), less exploration of the open arms and lower activity in the EPM, reduced social behaviour in the

(34)

social interaction test (Louvart et al. 2005), higher ASR (Milde et al. 2003), more immobility in the sudden silence test (Murison and Overmier 1998; van Dijken et al.

1992a; Van Dijken et al. 1992b; Van Dijken et al. 1992c) and short-lasting lower preference for a sweet solution (van Dijken et al. 1992a). Long-lasting behavioural changes induced by brief IFS are sensitive to treatment with (putative) anxiolytic agents, whereas no beneficial effect of antidepressant drugs is reported (Van Dijken et al. 1992b). In a previous study in our laboratory the diathesis stress hypothesis connected to PTSD and the HPA-axis was investigated. As expected, only rats with lower levels of corticosterone prior to brief IFS showed higher ASR (Milde et al.

2003). These data indicate that the behavioural consequences of a stressor may be related to pre-stressor levels of HPA-activity. Thus it has been argued that the brief IFS model is an animal model of anxiety, and more precisely PTSD (Stam 2007).

The predator stress model utilizes a stressor that is more naturalistic than footshock. In this model, rats are exposed to predator odour or threatened by a predator like a cat (but not physically attacked). Rats exposed to predator stress show anxiety-like behaviour in the following days and weeks, for instance increased ASR, and reductions in sexual behaviour, social interaction, weight gain and open arm activity in the EPM (Blanchard et al. 2003; Stam 2007). Anxiolytic and potentially anxiolytic drugs have been shown to modulate the elicited changes (Blanchard et al.

2003). Thus the predator stress model is known as an animal model of anxiety.

Chronic stress. The first chronic stress model of depression was developed by Katz (1981), where rats were exposed to several relatively severe unpredictable stressors. The harsh stressors applied raised ethical issues, leading to the development of the Chronic Mild Stress (CMS) model by Willner and colleagues (1987). In the CMS model rodents are repeatedly exposed to a set of various mild stressors across several weeks, supposedly mimicking the mild stressors humans are exposed to in everyday life (daily hassles). Rats exposed to CMS develop depression-like symptoms, for example sleep alterations, decreased sexual behaviour, weight loss, altered locomotor activity in the OF, decreased exploration, increased immobility in the forced swim test and reduced preference for sweet solutions (Grønli et al. 2004;

Grønli et al. 2005; Willner 2005; Yan et al. 2010). The effects can be reversed by

(35)

chronic treatment with antidepressant drugs and electroconvulsive shocks, while a number of anxiolytic drugs have no effect in the CMS model (Vollmayr and Henn 2003). Thus the CMS model is known as an animal model of depression.

Early life stress models include prenatal stress and maternal separation. In the prenatal stress model, the mother is for instance exposed to restraint stress, producing depression-like symptoms in the offspring. In the maternal separation model, rat pups are deprived of maternal care. The length and numbers of the separations have been shown to differently affect the rat pups. Brief maternal separation (e.g. 10 minutes per day) is shown to increase the resistance to stress in adulthood. Long-term maternal separation (e.g. 3 hours per day) may induce abnormal maternal behaviour (neglect) (Meaney et al. 1985), and may produce depression-like and anxiety-like symptoms in the pups that last into adulthood. Sleep alterations, elevated glucocorticoids response to stress, vulnerability to learned helplessness and ethanol self-administration, increased locomotion and decreased open arm activity in the EPM are seen after long-term maternal separation (Huot et al. 2001; Mrdalj et al.

2013; Nestler et al. 2002). Environmental enrichment and antidepressant drugs (Paroxetine, also used as anxiolytic) have been shown to reverse the depression-like and anxiety-like symptoms (Francis et al. 2002; Huot et al. 2001), indicating that early life stress is an animal model of both anxiety and depression.

The animal model of social defeat (SD) is also suggested as an animal model of both anxiety and depression, and is used in this thesis.

1.7 The animal model of social defeat

One of the main sources of stress in human life is of a social nature, like low ranking in the social hierarchy (Wood AM et al. 2012). Social defeat is associated with affective disorders i.e. anxiety and depression (Taylor et al. 2011). Studies in humans most often focus on school bullying and work harassment (Bjorkqvist 2001), and may be associated with traumatic events like violence and assault (Krug et al.

2002).

The animal model of social defeat (SD), most often using male rodents, is based on the resident-intruder paradigm first introduced by Ginsburg and Alle (1942).

(36)

A male rat intruder is placed in the territory of a bigger, older and more aggressive male resident rat. The intruder is attacked and defeated as indicated by fleeing, freezing and submissive behaviour (see Figure 1a,b). Behaviour and physiology are studied in the defeated intruder at different time intervals after stress exposure. Male rats are used in the social defeat model, as female rats do not normally show this aggressive territorial behaviour.

Figure 1. The white SD rat was (a) introduced, (b) defeated and further (c) exposed to the brown aggressive dominant rat. Total exposure to SD was 1 hour on one or two consecutive days.

Several variations of the SD model have been used. The nature of the social conflict may vary between only ‘physical attack’, and both ‘physical attack’ and

‘threat of attack’. The ‘physical attack’ phase is when the intruder is exposed to the resident and attacked. After being defeated, the intruder may be physically separated from the resident, protected from repeated attacks and potential injuries, but still being under ‘threat of attack’ by having auditory, visual and olfactory contact with the resident (see Figure 1c). This time under threat of attack is known to be highly stressful (Tornatzky and Miczek 1994). The number of exposures to SD varies from a single exposure to daily exposures, lasting for minutes, hours or even for weeks. In the present project, the main stressor for the intruder was SD for a total of 1 hour, including ‘physical attack’, subordination and further exposure by ‘threat of attack’

protected by a wire mesh cage. The intruders were exposed to SD on one or two consecutive days, respectively single SD and double SD.

Single or double exposure to SD in rodents has been shown to induce acute, short-lasting and long-lasting changes on behavioural, physiological and neuroendocrine parameters. Some effects of SD may be evident both acutely and last for days and weeks, and some may not be present acutely, but develop over time. In

a) b) c)

(37)

the present project the effects of single and double SD were examined on day 1 and up to day 24 after defeat.

1.7.1 Effects of single or double social defeat in rats

Acute effects of SD are shown during the social interaction and return to baseline during the hours after the termination of stress. These effects include increased corticosterone, ACTH, noradrenaline and adrenaline levels, increased core body temperature and increased heart rate (Heinrichs et al. 1994; Koolhaas et al.

1997b; Sgoifo et al. 1996; Tornatzky and Miczek 1994). Such acute effects were not studied in the present project.

In the OF test, rats exposed to single SD have decreased activity day 1, 2 and 7 after defeat (Meerlo et al. 1996a). In the same study a long-term effect appeared 28 days after defeat, as single SD rats showed increased latency for moving from the centre (where initially placed) to the periphery, compared to controls. Single SD has been shown to induce decreased central activity compared to controls when tested 7 days after defeat (Kavushansky et al. 2009). In a free exploration OF test, individually housed (compared to co-housed) single SD rats showed a longer latency to leave the home cage and less activity in the peripheral zone 21 days after defeat (Ruis et al. 1999). The latter effect was also seen on day 2 after defeat. However, effects of single SD on OF behaviour are not consistent, as some studies have shown a lack of short-term and/or long-term effects (Carnevali et al. 2012, (day 9 and day 21 after SD); Kavushansky et al. 2009, (day 1 after SD)). In the EPM test, studies have reported reduced percentage of time spent on open arms acutely after single SD (Heinrichs et al. 1992), an effect reversed by a CRH antagonist and by the anxiolytic midazolam. Another study showed increased latency to enter an open arm, reduced entries and time spent on open arms in the EPM 7 and 21 days after single SD (Carnevali et al. 2012). Reduced time spent on open arms has also been seen 14 days after single SD if the rats were housed individually and not in groups (Nakayasu and Ishii 2008; Ruis et al. 1999). However, there has also been reported unchanged activity on open arms or total activity in the EPM both 1 and 7 days after single SD (Kavushansky et al. 2009). Decreased preference for sucrose has been shown to develop day 22 after single SD (Carnevali et al. 2012), but also no preference for a

(38)

sweet solution have been reported (Meerlo et al. 1996b).The effect of single SD on ASR and sexual behaviour has not previously been reported. Sleep in SD rats has previously only been studied acutely after single SD. One study showed increased SWA during the active (dark) phase immediately following defeat, an effect that gradually vanished during the following inactive phase (Meerlo et al. 1997). Another study showed that SD rats kept awake by gentle handling after defeat, compared to controls kept awake by gentle handling during the same period, had a higher increase in SWA during NREM sleep (Meerlo et al. 2001). Responses of the HPA-axis i.e.

corticosterone and ACTH are acutely increased after SD (Heinrichs et al. 1994;

Koolhaas et al. 1997b; Sgoifo et al. 1996). Following single SD and a set of behavioural tests (OF, EPM and forced swim test), higher corticosterone levels have been seen compared to controls on day 7 after defeat, an effect that was not present day 1 (Kavushansky et al. 2009).

The predictive validity of single SD has previously been tested. Fluoxetine (antidepressant and anxiolytic drug) has shown to reverse body weight loss, reduced food intake and anxiety-like behaviour in the EPM (Berton et al. 1999).

Clomipramine (antidepressant and anxiolytic drug) has reversed anxiety-like behaviour (immobility) in the sudden silence test (Koolhaas et al. 1990). Reduced locomotion in the OF has been reversed by sleep deprivation (Meerlo et al. 1996a).

Anxiety-like behaviour in the EPM has been counteracted by a CRH antagonist and by the acutely anxiety reducing agent midazolam (Heinrichs et al. 1992).

To sum up, following single SD, rats in the OF normally show decreased total activity, decreased central activity and increased initial latency to move to a new sector. In the EPM they show reduced time on open arms and increased latency to enter an open arm. Single SD rats show reduced sucrose preference, and acutely increased SWA during sleep. These effects may be seen up to 4 weeks after SD, and may be associated with both anxiety-like and depression-like symptoms. Single SD may thus be regarded as an animal model of both anxiety and depression.

The effects of double SD have only been considered in a few studies. One study has shown that, compared to controls, both single and double SD induce reduced social interaction with a non-aggressive opponent, reduced total activity in

Consequences of social defeat stress for behaviour and sleep. Short-term and long-term assessments in rats